4.1 Introduction

Increasing demand for wireless communications to support deployed, military operations results in increased energy requirements for these missions. The appetite for greater connectivity on the modern battlefield, to support “net-centric” operations, only adds to energy demands for communications. In most cases, expeditionary command and control nodes cannot rely on existing infrastructure and must provide their own power. Diesel generators generally provide on-site power. Vulnerable, costly, ground-based logistics convoys supply the fuel.

Currently, military wireless communications (MWC) employ dedicated, stand-alone radio networks with each terminal having its own power supply and processors. In contrast, the industry has adopted an enterprise wireless communications (EWC) architecture for cellular base stations [1, 2]. The US Army and Navy have established research and development projects to investigate the EWC architecture for expeditionary networks. However, the extant military research focuses on providing connectivity while failing to address the potential power savings of EWC architecture [3]. This paper demonstrates replacement of the current expeditionary MWC architecture with an EWC architecture of software-defined radios (SDRs) has the potential to reduce overall power consumption by reducing powered-equipment redundancies such as processors and power supplies. This paper contributes to the literature on EWC architecture by conducting analysis to estimate the potential power reduction resulting from adoption of an EWC architecture for expeditionary, military operations.

4.1.1 Current Military Wireless Communications (MWC) Architecture

Under the current MWC architecture, an expeditionary node employs numerous, stand-alone, half-duplex radios. Each radio (or SDR) corresponds to a command and control network and serves a dedicated role or function. High-frequency radio networks are employed for over-the-horizon communications, for example, ship-to-shore communications. Very-high-frequency radio networks perform medium-range communications, for example, ground-tactical communications or convoy control. And, ultrahigh-frequency radio networks perform line-of-site communications, for example, air-to-ground communications. Figure 4.1 is a schematic representation of the current MWC architecture highlighting the hardware redundancies inherent in the architecture. The schematic also shows major components, along with notation, used for power modeling and simulation.

Fig. 4.1
figure 1

Schematic of software-defined radios (SDRs) employed using the current military wireless communications (MWC) architecture. Adapted from Fernando [4]

Full size image

To simplify power modeling and simulation, four general radio system types are defined based on their ability to perform power management. In general, and by design, current short-range radio systems use older technology and cannot perform power management. Medium-range tactical systems can perform limited power management. Long-range tactical systems employ newer technology, can perform power management, and use a medium complexity waveform. Finally, long-range tactical anti-jam systems employ newer technology, can perform power management, and use a complex anti-jam waveform.

4.1.2 Technological Advancements Enabling Adoption of an Enterprise Wireless Communications Architecture

Technologies have matured such that the realization of an EWC architecture for expeditionary applications is possible. Key enabling technologies include software-defined radios (SDRs), enterprise processing systems, modular open systems architectures (MOSA), and high-speed fiber optic serial interfaces.

The introduction of analog-to-digital converters and digital-to-analog converters and performance improvements in general-purpose processing brought about SDRs [5, 6]. Early digital radios allowed the use of simple waveforms in specialized digital subsystems to implement modulators and demodulators. Digital radios modulate digital information into an analog signal for transmission and digitize demodulated incoming analog signals. In the 1990s, general-purpose processor technology improved enough to allow a wide variety of waveforms to run. This innovation gave rise to the SDR [6]. Today’s SDRs can implement complex modulation and demodulation algorithms in software to increase the amount of digital information transmitted and received at any given frequency and time.

Advancements in processing performance and high-speed processing node connectivity have brought about an enormous increase in processing capability [7]. Processing nodes now have multiple processors with multiple processing cores that can process more data, faster than ever before. Moreover, processing capability scales up with the addition of processing nodes that communicate with other processing nodes at very high speeds. Interconnected processing nodes form a high-performance computing environment called a cluster. A large cluster is the technology behind cloud computing. A cluster, or enterprise processing system, can process multiple waveforms simultaneously [7].

MOSA provides the mechanism to use interoperable waveform software modules in an enterprise environment. It can also provide an abstraction between the hardware and software to allow plug and play capability for any waveform to run within an enterprise. The Department of Defense’s Software Communications Architecture (SCA) is a specific example of MOSA that creates a highly modularized architecture that detaches the waveform application from the underlying hardware platform [8]. SCA simplifies the implementation of multiple waveforms in SDRs.

Finally, the enterprise server, implementing the SCA , needs to transmit and receive digital signals at very high speeds. The digital signals, which require more than 10 Gbps of data throughput, would be too expensive to implement via parallel interfaces. The latest serial ports known as field programmable gate arrays are implemented using fiber optics and have this threshold. Furthermore, industry leaders in this technology have announced bandwidths of 1 Tbps over dual-mode serial ports [9].

Figure 4.2 is a schematic of SDRs employed using an EWC architecture. The major components are radio head modules (RHM), an enterprise server, and terminals. The RHM translates radio frequency (RF) signals to digital signals and vice versa. The enterprise server implements the software functions of a SDR. The enterprise server is connected to each RHM by high-speed fiber optics. The server modulates information for transmission, and the RHM synthesizes the modulated data into signals for the antenna to radiate. Likewise, the antenna receives the signals, the RHM digitizes the signals, and the enterprise server demodulates the digitized radio frequencies. The terminals execute applications to perform higher-level functions like user interfaces, integrated displays, and maintenance functions. In the EWC architecture, RHMs (a single RHM can replace several radios) replace the numerous, stand-alone, dedicated half-duplex radios of the current MWC architecture. In the figure, four RHMs are shown to depict the four general radio system types.

Fig. 4.2
figure 2

Schematic of multiple software-defined radios (SDRs ) employed using an enterprise wireless communications (EWC) architecture. Adapted from Fernando [4]

Full size image

The EWC architecture has all the advantages of an enterprise system [6], such as improved reliability, maintainability, and affordability. An enterprise architecture allows for easier and inexpensive upgrade of the enterprise server when performance and higher-efficiency processors become available. Moreover, the software is portable from the older enterprise system to the next-generation enterprise system [6]. Currently, expeditionary nodes employ a collection of disparate, dedicated communication systems, which have redundant power and processing systems. An enterprise system incorporates redundant processing subsystems that remain powered off when not required and could thereby reduce the power requirement of the current architecture while providing the same command and control capabilities.

4.2 Modelling and Simulation

4.2.1 Power Consumption Modeling

Power consumption is the amount of energy used for a given time interval, and total system power consumption is the fundamental comparison for this work. Power consumption is a function of instantaneous power, power modes, duty cycle, and radio system type.

Instantaneous power consumption is the power used by a system component at an instant in time. Equation (4.1) expresses the total instantaneous power of each component of an SDR (with the subscripts correspond to each component, left to right and top to bottom in Fig. 4.1: modulator, demodulator, controller, digital signal processor, digital-analog converter, analog-digital converter, transmit tuner, receive tuner, transmit filter, receive filter, transmit amplifier, receive amplifier, and digitally controlled switch, respectively).

$$ {\displaystyle \begin{array}{l}{P}_{\mathrm{total}}\left[W\right]={P}_{\mathrm{MOD}}+{P}_{\mathrm{DEM}}+{P}_{\mathrm{CON}}+{P}_{\mathrm{DSP}}+{P}_{\mathrm{DAC}}+{P}_{\mathrm{ADC}}\\ {}\kern4em +{P}_{\mathrm{TXT}}+{P}_{\mathrm{RXT}}+{P}_{\mathrm{TXF}}+{P}_{\mathrm{RXF}}+{P}_{\mathrm{TXA}}+{P}_{\mathrm{TXR}}+{P}_{\mathrm{DCS}}\end{array}} $$
(4.1)

For ease of simulation, we chose a differential approach, to obtain the following expression for power consumption as a function of instantaneous power:

$$ {P}_{\mathrm{con}}\left[W\cdot s\right]={P}_{\mathrm{total}}\Delta t $$
(4.2)

Since radios are half-duplex, transmit components are not powered during receive (and vice versa). More modern radios can also do power management. We refer to these as power modes and define four: sleep, standby, transmit, and receive. The amount of time that the HRM is transmitting or receiving is the duty cycle. To simplify modeling, we defined three duty cycles: light, medium, and heavy usage. Finally, power requirements vary depending on which of the four general radio types are being used (due to power management capabilities). To implement these considerations mathematically, a constant D (for duty cycle), is used to modify Eq. (4.2). Equation (4.3), coupled with tabulated power values, can be used to calculate power consumption for a system of SDRs.

$$ {P}_{\mathrm{totalcon}}\left[W\cdot s\right]={P}_{\mathrm{con}}D $$
(4.3)

4.2.2 Consumption Simulation

Power consumption simulation requires instantaneous power and duty cycle values with which to evaluate Eq. (4.3). Table 4.1 summarizes duty cycle profiles. Tables 4.2 and 4.3 summarize the instantaneous power values used for this work. The component differences in Tables 4.2 and 4.3 are architecture related, namely, power is shared in EWC but not under the current MWC architecture. The component power values and power supply efficiencies were found in the literature and verified against publicly available technical specification [1, 10,11,12,13]. Subject matter experts then validated both the component power requirements and duty cycle values [10]. Simulation was conducted with via spreadsheet. Interested readers are directed to Fernando [4] for a more nuanced discussion of these considerations and implementation.

Table 4.1 Duty cycles used for this work
Full size table
Table 4.2 Mode-based instantaneous component power requirements for SDRs operating under the current MWC architecture [1, 10,11,12,13]
Full size table
Table 4.3 Mode-based instantaneous component power requirements for SDRs operating under an EWC architecture [1, 10,11,12,13]
Full size table

4.3 Results

The intent of modeling and simulations was to determine the potential power savings that could be realized from replacing the current MWC architecture with an EWC architecture. We used data from Table 4.1 to evaluate the current MWC architecture with Eq. (4.3) for four SDRs operating in parallel, one for each of the four general radio types, operating for a week at light, medium, and heavy usage, respectively. Figure 4.1 shows this architecture. This time scale is long enough for duty cycles, like those defined, to emerge from real-world operations. We used data in Table 4.2 to evaluate the EWC architecture with Eq. (4.3) for a single SDR with four HRMs, one HRM for each of the four general radio types, operating for a week at light, medium, and heavy usage, respectively. Figure 4.2 shows this architecture. Table 4.4 summarizes simulation results.

Table 4.4 Summary of weekly power consumption simulations for MWC and EWC architectures for expeditionary command and control
Full size table

4.4 Architecture Power Consumption Discussion

The simulation results in Table 4.4 and comparison of MWC and EWC power usage values reveals reduced total energy usage for an EWC architecture regardless of scenario. As one might expect, the degree of power-energy reduction varies depending on radio system type and duty cycle.

Percent reduction is the energy reduction that could be realized by replacing the current MWC architecture with a EWC architecture. These values were calculated using the equation shown in Table 4.4. These values yield a lower and upper power reduction bound of 6% and 31%, depending on the radio type. Average power reduction is the average of the four percent reduction values.

The simulation estimates an average energy reduction of 15% across the entire simulation matrix. Restricting calculations to the medium usage numbers alone, with the assumption that this is the most probable operational scenario, an 11% power reduction is expected. These results suggest adoption of an expeditionary EWC architecture for SDRs would result in a meaningful reduction in energy requirements for expeditionary nodes without compromising requisite operational command and control capabilities.

4.5 Conclusions and Future Work

This paper presents a straightforward modeling, simulation, and power analysis demonstrating that an expeditionary EWC architecture would reduce operational energy requirements when compared to the current MWC architecture. This gain is, in part, due to the sharing and full utilization of power supplies and processing. In general, the underutilization of processing reduces efficiency [12]. However, modeling processor utilization is complicated due to the randomness of waveform processing occurrence. Additionally, waveform-specific processing specifications, in instructions per second (IPS), were unobtainable and would vary for different types of processors, making it necessary to test actual hardware and implement actual waveforms in software. Nonetheless, the simple approach used for this work likely provides a good first approximation for potential power reduction.

A follow-on proof-of-concept effort is necessary to prove these results with actual hardware. Actual hardware implementation would provide absolute data and determine any added power savings from enterprise processing. Real systems would also demonstrate possible emergent capabilities in network management, dynamic spectrum allocation, coalition interoperability, and electronic warfare. Modifying the current model and performing sensitivity analysis could potentially bind the impact of waveform randomness or IPS differences and inform the hardware test regimen.

The EWC architecture concept is already a reality in the commercial world. The cellular industry is driving future efficiency enhancements of RHMs, which include efficient technologies in power supplies and amplifiers. The push for efficiency in the enterprise computing industry is already in crescendo. As this paper demonstrates, adoption by the military can result in noticeable power savings, which is a valuable benefit to expeditionary operations where fuel must often be transported to support operations.